Inflation Cooled. The 30-Year Didn't.

First, the Numbers — Straight Off the Wire

The July Consumer Price Index landed this morning. Prices rose 0.1% from June, running at a 3.4% annual rate, down from 3.5%. Core prices — stripping out food and energy — matched the consensus estimate too. Not hot. Not cold. Exactly what the market had already priced.

The bond market took it well at the short end. The 10-Year Treasury — the benchmark that actually drives your mortgage — eased about three basis points to roughly 4.654%. The 2-Year came off four, to around 4.178%. Both moved the way you would expect on a cooling inflation print.

The 30-Year ticked down too, by not quite two basis points, to about 5.217%. On the day, that is a decline. Hold that number still for a second, though, because the day is the wrong frame.

Market data as reported the morning of August 12, 2026. Not a quote, not an offer, not a prediction. These numbers move continuously and may have moved by the time you read this.

The Gap I Asked You to Watch

On July 30, in No. 5 of this series, I ended with a specific instruction rather than a forecast: watch the gap between the 10-Year and the 30-Year. That day the 30-Year printed 5.20%, which the coverage called a nineteen-year high, and I said the pressure behind it “does not stay politely in its own lane forever.”

Two weeks later, here is the read:

So the honest sentence is not “the long end sold off today.” It did not. The honest sentence is this: across two weeks that included a cooling inflation report and a weak jobs report, the short end came down and the long end went up. Two pieces of news that should have helped both ends of the curve only helped one.

I said in No. 5 that I would not hide the argument against my own position in a parenthesis. This is that argument arriving on schedule. A 30-Year above a nineteen-year high, after data that should have pulled it down, is the bond market saying it is not finished worrying about long-run inflation — whatever the July CPI headline said.

What That Does and Does Not Mean for You

It does not mean your mortgage rate follows the 30-Year bond.It does not. I said so plainly two weeks ago and it is still true: your 30-year fixed prices off the 10-Year and mortgage-backed bonds, not the 30-year Treasury. Anyone flattening this into “rates hit a nineteen-year high” is telling you something false.

What it means is narrower and more useful. The part of the curve that prices your loan is behaving. The part that prices long-run inflation risk is not. Those two can disagree for a while. Historically they do not disagree forever.

Let Me Argue Against My Own Urgency First

Today's report was, on its face, good news for borrowers. Inflation cooled. Yields fell. A weaker jobs report last Friday already had investors leaning toward the Fed sitting still in September. If you want a case for waiting, it is not a stupid one, and I am not going to pretend otherwise to move you.

There is also a real cost to acting badly. Refinancing into a rate that barely beats your current one, and paying closing costs to do it, is how people lose money while feeling productive. A refinance that does not clear its own costs in a sensible number of months is not a win. It is a fee.

So no, I am not telling you the sky is falling. I am telling you what is on the calendar.

“In Line” Is Not a Green Light. It Is a Deferral.

Here is what an as-expected print actually does: it changes nothing, so it settles nothing. The market did not learn the answer this morning. It learned that the answer arrives later.

Ian Lyngen, who runs U.S. rates strategy at BMO Capital Markets, put it about as plainly as a strategist ever does: the release “leaves the path open for the Fed to pause in September, but is by no means definitive,” and “the September decision now comes down to the August payrolls and CPI combination.”

Read that as a borrower rather than a bond trader and it says something specific. The decision that prices your mortgage this autumn has been handed to a small number of reports that have not been written yet.

The Detail Almost Nobody Is Repeating

At the July meeting, with the target range at 3.50%–3.75%, three members of the committee dissented. They voted to raise.

That is not a rumour or a forecast. It is a recorded vote. And it is the single most useful fact in this entire report, because it tells you the direction of the risk is genuinely two-sided. The headline says inflation cooled. The vote says three people in the room were not convinced.

I have no idea which side is right. Neither does anyone quoting you a September forecast with a straight face. What I know is that a two-sided risk is exactly the condition under which being un-decided is itself a position — and most people holding that position have not noticed they are holding it.

What This Actually Means, Depending on Who You Are

The Only Number That Matters Is Not in This Article

Everything above is public. The CPI print, the yields, the dissent, the calendar — you can verify all of it, and you should.

What none of it tells you is whether the math works on your loan. A 3.4% inflation reading is not a reason to refinance. Neither is a 4.654% 10-Year. The reason to refinance, or not to, is the gap between what you are paying and what you could be paying, measured against what it costs to make the change — and that number belongs to you, not to the market.

I am not going to guess it in a blog post. But you can have it in about two minutes.

The Watch Continues — On the Record

This is entry No. 6 in The 10-Year Watch — my public, on-the-record read of this market, published before the fact so you can judge it after. No. 1 called 4.67%. No. 3 watched it tested. No. 4 said protection was cheap while it was still optional. No. 5 kept score at 4.671%.

This one is the first read on the gap No. 5 asked you to watch, and it came back split: the end that prices your loan eased, the end that prices long-run inflation risk did not. I will post what actually happens after August CPI, including the version where the long end settles back down, the Fed sits, and this entry reads like a lot of throat-clearing about a quiet Wednesday. That outcome is entirely possible. And one call holding for thirteen weeks is not a track record — it is one call, still holding. Publishing the scorecard either way is the price of publishing the call at all.

Related: Keeping Score at 4.671% · Locking Isn't a Prediction · The 4.67% Ceiling · Points & Lender Credits · Construction-to-Permanent Loans · VA Loans · Physician Loans